User Intention based Personalized Search : HPS(Hierarchical Phrase Serch)

Personalized search has recently got significant attention in the web search. Accordingly, user’s intention of search is very important information to retrieval information in aspect of personalized Search. Although many personalized search and strategies have been proposed, the majority of web users are difficulty in retrieval information corresponding their search intention. In this paper, we present personalized search based on User Intention through the HPVM(Hierarchical Phrase Vector Model) to solve these problems using and machine learning methodology. Users can navigate through the prior user’s intention by their own needs. This is especially useful for various meaning and poor queries. By analyzing the results, we find out that there is an unique representation of user intention under different queries, contexts and users. Furthermore, we can find out that this knowledge is very important in improving personalized information retrieval performance by filtering the results, recommending a new query, and distinguishing user’s characteristics. With this approach, search engines can provide more predictive information for Web searchers. Based on this approach, we developed a personalized search engine, HPS (Hierarchical Phrase Search). Key-Words: Personalization, Information retrieval, Hierarchical phrase search, Support vector machine, Machine learning, Relevance feedback